AI for Social Empowerment_ Driving Change and Inclusion

20 Feb 2026 11:00h - 12:00h

Session at a glance

Summary

This discussion focused on the impact of artificial intelligence on labor markets and employment, featuring perspectives from technology, research, and policy experts. The conversation was moderated by Anurag Behar from the Azeem Premji Foundation, with panelists including Sabina Dewan from Just Jobs Network, Julie Delahanty from IDRC Canada, and Sandhya Ramachandran Arun from Wipro Limited.

Sabina Dewan argued that AI’s impact on jobs is already evident, citing private company admissions of 30-40% productivity gains that translate to workforce cuts, despite public denials of job disruptions. She emphasized that major tech companies are already laying off thousands of workers and warned that AI represents unprecedented disruption, noting that even AI pioneers are raising alarm bells. Dewan stressed the urgency of implementing regulations, social institutions, and protective systems rather than waiting for more empirical evidence.

Sandhya Ramachandran Arun, representing the technology sector, presented a more optimistic view, explaining that while AI automates coding tasks, it elevates workers to more strategic roles requiring human oversight, creativity, and business understanding. She argued that the industry has been transforming job roles for years, focusing on skills like adaptability and communication rather than just technical coding abilities.

Julie Delahanty highlighted the importance of strong institutions, human-centric AI development, and evidence-based policymaking. She emphasized that responsible AI governance requires co-creation with workers and communities, along with robust research to understand real-world impacts on labor markets.

The discussion concluded with agreement that AI represents a transformation as significant as nuclear technology, requiring immediate action on regulation, education reform, and social protection systems to manage the inevitable disruption while preserving jobs requiring human wisdom, empathy, and understanding.

Keypoints

Major Discussion Points:

– AI’s Current Impact on Jobs is Already Happening: Sabina Dewan argues that contrary to claims that AI’s job impact is “yet to unfold,” companies are already experiencing 30-40% productivity gains leading to workforce cuts, and major tech companies have been laying off thousands of workers, though they don’t publicly attribute this directly to AI.

– The Nature of Work Transformation vs. Job Displacement: The panelists debate whether AI will primarily displace jobs or transform how work is done. Sandhya from Wipro emphasizes that coding jobs are evolving into more managerial roles overseeing AI agents, while maintaining that human wisdom, creativity, and strategic thinking remain irreplaceable.

– Urgent Need for Regulatory and Institutional Response: There’s strong consensus that governments cannot afford to “wait and see” but must immediately develop comprehensive policy frameworks including competition policy, tax reform, labor regulations, universal social protection systems, and improved education/skills training to manage AI’s disruptive effects.

– Global South and Informal Economy Vulnerabilities: The discussion highlights particular concerns for countries like India where only 9-10% of employment is formal sector, meaning AI disruption of these “good jobs” could have cascading economic effects, while the majority of workers in precarious informal employment lack safety nets.

– AI as Retail-Level Societal Transformation: The conversation concludes with recognition that AI represents a uniquely pervasive disruption compared to previous technologies like nuclear power, affecting every individual through education, cognitive development, and daily work practices in ways that are extremely difficult to regulate.

Overall Purpose:

The discussion aimed to examine AI’s real-world impact on labor markets, moving beyond abstract speculation to assess current evidence of job displacement and transformation, while exploring policy responses and institutional changes needed to manage this technological disruption responsibly, particularly in developing country contexts.

Overall Tone:

The discussion began with a provocative, urgent tone set by Sabina Dewan challenging the notion that AI’s impact is still unknown. The tone evolved through cautious optimism from the tech industry representative, balanced pragmatism from the development research perspective, and ultimately converged on serious concern and urgency. By the end, there was a sobering recognition of AI as potentially the most disruptive technology in human history, requiring immediate, comprehensive action despite the complexity of the challenge.

Speakers

Speakers from the provided list:

– Sabina Dewan – Labor market expert and researcher, representing Just Jobs Network

– Anurag Behar – Chief Executive Officer of the Azeem Premji Foundation, moderator of the discussion, works in education sector (runs three universities and works with over 100,000 teachers)

– Sandhya Ramachandran Arun – Chief Technology Officer of Wipro Limited, technology industry representative

– Julie Delahanty – President of IDRC Canada (International Development Research Centre), focuses on AI governance and development research

Additional speakers:

None identified – all speakers mentioned in the transcript are included in the provided speakers names list.

Full session report

This panel discussion examined the profound implications of artificial intelligence on labour markets and employment, bringing together diverse perspectives from technology, research, policy, and education sectors. Moderated by Anurag Behar from the Azeem Premji Foundation, the conversation featured Sabina Dewan from Just Jobs Network, Julie Delahanty from IDRC Canada, and Sandhya Ramachandran Arun from Wipro Limited. Notably, Behar disclosed a significant conflict of interest at the outset, explaining that the Azeem Premji Foundation owns approximately 70% of Wipro, creating tension between tech company interests and his responsibility to vulnerable populations.

Setting the Provocative Context

Sabina Dewan opened the discussion by framing it as a “provocative context-setting,” challenging the prevailing narrative that AI’s employment impact remains uncertain or distant. She highlighted the unprecedented nature of current concerns by noting that AI pioneers themselves—including Geoffrey Hinton, Stuart Russell, and Dario Amodei—are raising alarm bells about the technology they helped create. This represents an extraordinary situation where the creators of a transformative technology are warning about its potential dangers.

Dewan pointed to the massive capital accumulation in the AI sector, exemplified by NVIDIA’s $5 trillion market capitalisation, as evidence of the technology’s transformative economic power. She argued that the employment impacts of AI are not hypothetical future concerns but present realities that companies are experiencing now, even if they are reluctant to acknowledge them publicly.

The Current Reality: Evidence of Immediate Impact

Drawing from extensive research, particularly in India, Dewan presented compelling evidence that companies are already experiencing significant workforce disruptions. Her research revealed a stark disconnect between public corporate messaging and private reality: whilst companies publicly maintain their role as job creators, private conversations reveal acknowledgements of 30-40% productivity gains from AI implementation, which directly translate to substantial workforce reductions.

This revelation challenged the common assumption that AI’s impact on employment remains speculative. Dewan emphasised that major technology companies have already laid off thousands of workers, though they rarely attribute these cuts directly to AI, instead citing macroeconomic conditions or other factors. However, she argued that AI represents a significant disruption layered atop existing challenges such as the pandemic, climate change, and trade shocks.

The evidence extends beyond mere job quantity to encompass quality of work. AI systems are already enabling increased surveillance of workers, influencing decisions about work allocation and entitlements, and facilitating algorithmic management in the gig economy. This creates situations where workers wronged by platforms have no recourse for redressal, as they are managed by algorithms rather than human supervisors.

Technology Sector Transformation: Adaptation Over Displacement

Representing the technology industry, Sandhya Ramachandran Arun offered a perspective focused on adaptation rather than simple displacement, whilst acknowledging the reality of significant change. She explained that the IT industry has been undergoing transformation for over fifteen years, consistently trying to communicate that software engineering roles encompass far more than mere coding—a message that has taken fifteen years to convey effectively.

The advent of AI has accelerated this evolution, with coding tasks increasingly handled by AI agents whilst human workers focus on higher-level responsibilities including business understanding, customer experience design, architecture, and security oversight. This transformation has fundamentally altered hiring practices within the technology sector, with companies now prioritising learnability, communication abilities, and adaptability over purely technical coding skills.

Arun described how junior developers are evolving into “AI managers” who oversee AI-generated code whilst ensuring it meets business requirements, security standards, and architectural principles. This represents an elevation of roles rather than simple displacement, requiring workers to develop new skills in managing AI systems whilst maintaining human oversight of critical decisions.

The transformation extends beyond the technology sector itself. In marketing, AI can generate high-quality visual, audio, and video content, but strategic planning, execution oversight, and return on investment analysis remain fundamentally human responsibilities. Similarly, in finance, whilst AI can handle much processing work, human wisdom remains essential for data interpretation and decision-making.

Governance Challenges: The Need for Institutional Frameworks

Julie Delahanty brought a global development perspective to the discussion, emphasising that effective AI governance requires robust institutional frameworks rather than merely technological solutions. Her experience with IDRC’s AI4D work across multiple countries highlighted the critical importance of strong regulatory institutions, labour protections, and research ecosystems in managing AI’s labour market impacts.

Delahanty stressed the need for human-centric AI development through co-creation with workers, communities, and employers. This approach aims to enhance job quality and productivity rather than simply increasing inequalities. She referenced work on a Global Index on Responsible AI that provides comparable data across 138 countries, with particular focus on labour protection and the right to work.

However, Delahanty acknowledged that the governance challenge is unprecedented. Unlike previous technological disruptions, AI’s pervasive nature makes it extremely difficult to regulate effectively. The technology penetrates every aspect of work and daily life, creating regulatory challenges that traditional frameworks struggle to address.

India’s Unique Vulnerabilities: The Formal-Informal Employment Nexus

The discussion took a particularly sharp focus when Behar posed a provocative question: given that only 9-10% of India’s employment is in the formal sector, why should the country be concerned about AI disruption affecting such a small proportion of workers?

Dewan’s response illuminated the cascading effects of formal sector job losses. She noted that 58% of India’s employment is now self-employment, highlighting the precarious nature of work for the majority of the population. She argued that India’s limited formal employment makes AI disruption even more devastating rather than less concerning. The country has so few “good jobs” that disrupting this small formal sector would push more workers into precarious informal employment, exacerbating existing inequalities.

The interconnected nature of formal and informal employment means that job losses in sectors like information technology create ripple effects throughout the economy. When IT workers in Bangalore lose their jobs, the restaurants, bars, housing, and automotive sectors that depend on their spending also suffer, creating cascading unemployment that extends far beyond the initial AI-driven displacement.

Furthermore, India’s workforce faces fundamental challenges in adapting to an AI-driven economy. Despite years of investment in skills training programmes, only 4.1% of the labour force acknowledges having any formal skills training. The quality of basic education remains poor, with many graduates unable to perform foundational reading and mathematics despite completing secondary education.

The Educational Crisis: Cognitive Decline and Learning Disruption

Perhaps the most alarming aspect of the discussion emerged when examining AI’s impact on education and human cognitive development. Behar, drawing from his experience running three universities and working with over 100,000 teachers, described AI as “attacking the very foundation of education” by encouraging the outsourcing of thinking processes.

The educational implications are immediate and severe. Teachers and students are increasingly relying on AI to perform cognitive tasks that are essential for learning and development. This has forced educational institutions to return to traditional assessment methods—paper-and-pencil examinations conducted in controlled environments—as the only way to ensure authentic evaluation of student capabilities.

Dewan connected this educational crisis to broader labour market concerns, citing research showing that “for the first time, the current generation of young people have shown cognitive decline.” This represents the first time in modern history that cognitive abilities are declining rather than improving across generations. Increased rates of depression, anxiety, and measurable cognitive decline create a troubling feedback loop: as humans become less cognitively capable, they become more vulnerable to replacement by AI systems that are becoming more efficient.

Comprehensive Policy Framework: Urgent Interventions Required

The panel reached strong consensus on the need for immediate, comprehensive policy responses rather than a “wait and see” approach. Dewan outlined a detailed framework for necessary interventions, emphasising that waiting for more empirical evidence would be “way too late” given the pace of change and the evidence already available.

The required policy responses span multiple domains:

Competition Policy and Antitrust: Essential to address the massive concentration of capital in technology companies and prevent monopolistic control over AI development.

Tax Policy Reforms: Including wealth taxes, transaction taxes, and corporate tax adjustments to address growing inequality and fund social protection systems.

Labour Law Reforms: Must address the changing nature of work, particularly in the gig economy where workers are increasingly classified as independent contractors rather than employees, leaving them without health insurance, job security, or other protections.

Universal Social Protection: Essential to enable workers to transition between occupations and sectors as AI disrupts traditional employment patterns.

Education and Skills System Overhaul: Investment requires fundamental reimagining rather than incremental improvement. Current approaches that attempt to train workers with poor foundational education for AI-related work are inadequate. The systems must meet people where they are, providing basic literacy and numeracy skills before attempting more advanced training.

Managing Unprecedented Transformation: The Nuclear Technology Analogy

A striking theme that emerged was the comparison between AI and nuclear technology in terms of transformative potential and risk. Arun noted that humanity has successfully managed nuclear technology despite disasters, suggesting that similar governance approaches could work for AI. However, Behar argued that AI presents even greater challenges because it represents a “retail transformation of humanity” that affects every individual in ways that nuclear technology never did.

This comparison highlighted both the potential for human agency in managing technological disruption and the unprecedented nature of AI’s pervasive influence. Unlike nuclear technology, which remained largely contained within specific industries and government control, AI penetrates education, work, social interaction, and cognitive processes in ways that make comprehensive regulation extremely difficult.

Human Elements in Future Work

Despite the serious concerns raised throughout the discussion, the panel identified important areas where human capabilities remain irreplaceable. Behar concluded that jobs requiring “wisdom, empathy, care, and human understanding” emerged as the most likely to survive AI disruption. These qualities—creativity, vision, foresight, and human-centricity—represent core competencies that AI cannot replicate.

This insight suggests that the future of work may increasingly centre on fundamentally human capabilities rather than technical skills that can be automated. However, this transition requires significant investment in developing these human qualities through education and training systems that currently underemphasise such skills in favour of technical competencies.

Conclusion: Navigating Transformation with Human Agency

This comprehensive examination of AI’s impact on labour markets revealed the complexity and urgency of the challenges facing policymakers, educators, and workers. The discussion moved beyond simple questions of job displacement to examine fundamental issues of work quality, human cognitive development, institutional capacity, and social protection.

The panel’s diverse perspectives converged on several key points: AI is already transforming work in significant ways, the impacts extend far beyond simple job losses to encompass work quality and human development, and immediate policy responses are essential rather than optional. The conversation highlighted the particular vulnerabilities of developing economies with large informal sectors and weak institutional frameworks, whilst recognising that even developed economies face unprecedented challenges in governing a technology that penetrates every aspect of human activity.

Ultimately, the discussion reinforced that AI represents not merely a technological shift but a fundamental transformation of human society that requires comprehensive, immediate, and coordinated responses across multiple domains of policy and institutional development. The stakes are high, the timeline is compressed, and the need for action is urgent, but human agency and institutional innovation remain the key determinants of whether AI’s transformation of work enhances human flourishing or exacerbates existing inequalities and vulnerabilities.

Session transcript

Sabina Dewan

say, you know, it’s yet to unfold. We don’t know what the impact is and it’s yet to unfold. I believe that that contention is actually largely untrue. And let me tell you why. When you talk to companies privately, publicly they will not own up to the potential job disruptions as a result of AI. And partly that is because many of the big companies actually are known to be formal job creators, right? And that is a very important part of their image and their contribution to economies and societies. But when you talk to them privately, in India especially, our research shows that they will own up to anywhere between 30 % to 40 % time saving, right, productivity gains, which then translates into significant workforce cuts.

We already have plenty of empirical evidence that suggests that… that AI systems are enabling surveillance, they’re influencing decisions about who gets work, when, and what entitlements people have access to. We also know that AI systems are grossly exacerbating inequality. If you just look at the market caps of some of the top technology companies, you know, NVIDIA’s $5 trillion market cap, right? So there’s a massive accumulation of capital that really, you know, capital share is growing and labor share of income is getting smaller and smaller. So I guess, you know, this discussion that talks about social empowerment, a key question in that is the question of the impact on jobs. And the question that I, you know, put out there is, so if you even buy the idea that we don’t know, that we don’t know what the impact is, what the impact is going to be.

Can we afford to just wait, right? Or do we need to take every action possible in terms of regulations, in terms of building social institutions, in terms of really working to build systems that can manage this inevitable evolution of AI, whether we like it or not. The last thing I’ll say is just, you know, yes, there have been technologies before. Yes, they’ve had their own forms of inclusion and exclusion. But at the end of the day, this is the first time where you have the very pioneers of that technology, Jeffrey Hinton, Stuart Russell, Dario Amadai, the very pioneers of the technology themselves are ringing alarm bells. And would we not be wise to heed them?

So with that, I hope, provocative context setting, I am really grateful. On behalf of the Just Jobs Network, again, with support from IDR. CNF CDO to welcome our really esteemed panelists. Mr. Anurag Bihar, who is the chief executive officer of the Azeem Premji Foundation, has very graciously agreed to chair this conversation, moderate the discussion. We have Dr. Julie Delhanti, who is the president of IDRC Canada. Thank you, Julie. And Ms. Sandhya Ramachandran Arun, who is the chief technology officer of Wipro Limited. Thank you so much for being here, Sandhya. So, Anurag, over to you.

Anurag Behar

Thank you. Thank you, Sabina. Good evening, everybody. Thank you. There’s so much. There’s so much investment going into AI. why is it going into a why is so much investment there in AI? We are in the fifth day of the AI summit. So this is like the 42nd kilometer of a marathon. Right? At this stage, such investment has to be justified by some monetization. And where is that monetization going to come from? It’s either going to come from productivity, which comes from labor reduction, or it is going to come from new products and services or both, a combination of both. That’s where it’s going to come from. Right? We will talk more about that. At this moment, my job is easy.

I’m going to just ask Sandhya, because she’s the representative of the technology world here really, that which way is this technology headed? And in very simple terms, what is she seeing its implications on jobs? I mean, what kind of jobs are going to get displaced, destroyed? And what kind of jobs are going to get created? and what’s the underlying dynamic because of which these jobs will be created and the jobs will be destroyed. So how does she see it in the world of technology? Let’s start with that.

Sandhya Ramachandran Arun

Sure, thank you so much. Thanks, Anurag, for the question. So as far as the tech industry is concerned, we are really witnessing a very huge impact of the AI evolution as a disruptor. We’ve had to revisit how job roles are created. We’ve had to revisit how talent has to be reskilled. And we have also revisited the responsibility, not just in terms of security, safety, but also in terms of what does it mean to our colleagues and our hiring. I think initially there was a huge amount of fear that we would not hire from colleges, which is now… despair because we’re broken. continues to hire from colleges, and so do our competitors. But the criteria for hiring has shifted to a more nuanced, a more calibrated way of looking at learnability, looking at whether a person communicates well, technical ideas, looking at whether a person is adaptable.

Because AI is a technology that is changing as we speak. So no one can claim to be an expert in AI and remain that way for the next five days, possibly, because there’s things that’s going on changing every day. With regard to our own talent, we have created role personas, and we have created very specific learning modules on how the role changes with AI. And everybody from the board to the CEO down to the youngest employee is going through a very calibrated learning process. And there is also a very… calibrated way in which services and ways of working are changing. So to that extent, we see a change. We are not seeing a displacement because most of the work that we do is consultative in nature, inspired of the market valuation erosion that we saw some time back because of a news from Anthropic and Palantir.

The insiders in the technology world were already aware of the transformative nature of these solutions coming up. And we have already been using these solutions significantly for over a year. So from a market sentiment point of view, possibly there was an erosion, but from a technology impact perspective, we have been bracing ourselves for the change and our journey of transformation continues.

Anurag Behar

I just have a follow -up on that, and then I’ll move to Julie. I’ll put it very, I mean, let’s say, a very, very simple, commonsensical question. Which is that, we are hearing about these tools where coding has become so much more easier, right? So, and this is not just about Wipro, it’s about the IT industry in general. So if coding is becoming so much easier, and 50 % or 70 % of coding can be done by these AI tools, then isn’t it inevitable that IT sector jobs will be lost? Or if there’s business or volume growth, much less hiring will happen. So that’s part one to my question. Part two is, if you move away from the IT world, and if you go to let’s say design and marketing, or, I mean, let’s say my world of the academy, the world of research, so many of research assistants and those of you who have used research assistants or work with research assistants, so much of that job is being done easily by AI.

So part one of my question, if coding is becoming so much more efficient, isn’t it inevitable jobs will be lost? so much hiding will not happen, whichever way. And aside from that, in the outside world, in other industries, what is it that you’re seeing?

Sandhya Ramachandran Arun

Sure. Let me just address the coding part of it. I think for over 15 years, the industry has been trying to explain to the outside world and as well as to the talent aspiring for careers with us that we do not have coding roles primarily. Coding is a very small task in what a software engineer does or a software developer does. There is the need to understand business outcomes. There’s a need to understand customer experience. There’s a need to understand architecture and what is a well -engineered code, right? So this is not new today. This has been in existence. I mean, I’ve been doing digital transformation for the last 15 years, and we’ve been trying to change how the world thinks about these roles.

Yes, the day is here when coding can be completely handed off to an AI agent. And that is indeed a fact, right? But the fact that supports the success of this code in business is really the ability to have a human oversee the design, the engineering, the architecture, the security, as well as delegating the coding work to an agent. So the role of a junior developer really becomes that of a little manager of AI, as opposed to saying, you’re displacing my job. The person’s actually going up if the person really is aware and aligns to what the organization needs in terms of figuring out what is required. And those are the trainings that are happening.

That’s what’s happening in terms of selection. We now have COEs inside engineering colleges where we are talking to universities about this as well. And what about other industries? whatever you’re seeing? So other industries we work with, there is a variation. So if you think about it, marketing, there’s a lot of work that gets offloaded. The strategy, the planning, the oversight on execution, the ROI on marketing still remains a strategic thinking job that remains with humans. But you can generate a lot of good quality visual, audio, and video content using AI today. And probably it’s making marketing a whole lot more efficient. Now, if you take finance, for example, again, a lot of processing gets taken over by AI, but it still needs a human to bring in wisdom in terms of how the data gets interpreted, how decisions are being made, and also to make sure that the AI aligns to human values in some sense.

So those kind of changes are happening in these functions. And that’s why I’m so excited to talk to you today. And I’m so excited to talk to you today. And I’m so excited to talk to you today. And I’m so excited to talk to you today. And I’m so excited to talk to you today. Industry -wise, there is a lot happening positive, I would say, in, say, healthcare, for example, even in banking, for example, where we are able to fight financial crimes a whole lot better. In healthcare, we are augmenting technicians, clinicians, and doctors with more intelligent input for decision -making. And while AI can make the decision, you don’t allow it to make

Anurag Behar

So, Sandhya, just put a pin on something that you said, and I’ll come back in the second round. You used the word human and wisdom. So just put a pin on that, and I’m going to come back to that in my second round. Julie, if Sandhya was less than as optimistic as she is, she wouldn’t be representing the tech world, you know. So one should expect that she’s as optimistic. she is. But what I wanted to ask you was that, you know, eventually, and, you know, from your vantage point, you know, you’re seeing how governments are dealing with this evolving situation, and not just an AI safety and, you know, all the other things, but particularly on labor markets.

So how can governments and institutions govern AI responsibly, such that any disruption in labor markets is sort of minimized or handled well, or the transition happens well? So let’s assume this picture that Sandhya has painted, that, of course, there’s something that’s going on, the reproductive, like she talked about the marketing and advising. So some people are going to lose jobs there. So what should government institutions do? How does one govern this situation, such that the benefits are maximized? And I’m talking particularly about labor. markets, not the other stuff, while harms are minimized.

Julie Delahanty

Yeah, thank you so much. I’m going to answer that question, but the last question just made me think about two things. One was, you know, I’m old enough to remember when computers first came around in the 70s and, you know, what we thought would happen with computers and the job losses that we anticipated. And, of course, we did lose jobs. There was a lot of labor disruption related to, you know, typing pools and different kinds of ways. But at the time, even home computers, nobody could even fathom what you would do with a home computer. The conversation then was that home computers would be used to develop recipes and that you’d have recipes because homes were only where homemakers were.

People couldn’t even, there’s such gendered ideas that people just could not understand what you would do with a home computer. So I think in the same way, some of what… is going to happen with AI in the labor market, we may not be able to anticipate just yet. So just as a reminder of where we came from with other important technologies. But when it comes to governance, I think the important issue is that it’s not really only about the technology, it’s really about institutions, it’s about workers, and it’s also about research. So when it comes to institutions, really without the kind of strong institutions in countries, regulatory institutions, labor institutions, strong research ecosystems that are able to really understand what’s happening in the labor market, I think it’s very difficult to end up having a strong regulation of what’s happening in the labor market.

So just those institutions are incredibly important to understanding where job losses might be, where biases might happen, and really investing in people and institutions is something that has to go hand in hand with our thinking around technologies. Another… Another area is around making sure that when we’re thinking about new technologies, that we’re making it very human -centric. And one of the things that the AI4D program does when we think, what do we mean by human -centric? It’s really about making sure that we’re co -creating new technologies with the co -creation of workers, of communities, of employers, so that we can understand how to enhance job quality, how to enhance productivity, rather than increasing inequalities or changing who benefits.

So really understanding who benefits, who’s going to face the kinds of disruptions is really important so that we’re not thinking about that as an afterthought. That we’re really shaping AI systems using that knowledge. And similarly… I think the importance of research in, and I’ll just give an example from our AI4D work is we’ve done a big research program with partners in sub -Saharan Africa that’s looking at, that’s collecting household data, firm level data, worker data, to understand what the real world impacts of AI are on labor markets. And it’s that kind of tracking, who’s going to benefit, understanding who’s going to be displaced, and how the tasks and skills are really changing that’s going to allow governments to better design and think about what kind of skills development they need, what kind of social protections they need, and how to support labor rights.

So really, I think growing AI responsibly doesn’t mean avoiding innovation or avoiding change, but it’s really about shaping AI so that it, it does strengthen labor markets and supports workers and creates more opportunities.

Anurag Behar

Thanks, Judy. Thank you so much. I’ll move to Sabina Sabina I mean since you are the labor market expert here amongst us and the researcher what is it that you see what is it that I mean there’s so much of news we have had this five days of this grand summer what is really going on what do we understand what we don’t understand in the context of the impact of AI on jobs how do you stack up

Sabina Dewan

so just a little tongue -in -cheek we go back to the 1600s we’d asked chat GPT then if Galileo was correct it would have said no way right so this technology you know for all the possibilities that it brings notwithstanding it is not just a technology we can’t just at AI as machine learning, large language models. It is a system, it is an instrument that is being utilized for social, political, and economic engineering. And my job is to look at the impact of that in labor markets. So if we limit ourselves just to the question of how many jobs will be lost, how many jobs will be gained, that’s A, not even an appropriate question.

Two, I agree with my fellow panelists that we don’t necessarily know what sort of new possibilities there might be. But what we do know, what we already see, is also something that Sundar talked about, which is the efficiency gains. And any time there are efficiency gains, there are layoffs. And please, you do the research, right? Like, I do my job. But look at the newspapers. Companies are laying off thousands of workers already. All the big tech companies have in recent years been laying off workers. Now, sure, they can say that this is a confluence of many factors. It’s not just AI, and most of them will not just ascribe it to AI. They might ascribe it to macroeconomic conditions, to the confluence of various other forces like the pandemic or trade shocks, all of which is true.

But AI is one really big disruption that comes on top of all the other disruptions, and there’s already plenty of evidence that is suggesting that these disruptions are not just changing the quantity of jobs in terms of how many companies are already laying off workers. Again, I mean, we’ve heard also projections from the, tech companies themselves, right, what the possible projections are. of disruptions and layoffs are going to be. But we also already have evidence of people being laid off. But then on top of that, I would say let’s look beyond just how many jobs are lost and how many jobs are gained to actually look at, I mean, take the gig economy, for example, and algorithmic management of gig workers.

That is a labor market issue. If a gig worker is wronged, the platform just, you know, they just get kicked off the platform. There’s no mechanism for redressal because it’s an algorithm that’s managing the worker. So who do you talk to? I mean, I can go on and on and on. Now, we might be separating out platforms from AI, but actually the algorithms are AI, and it’s embedded in a platform economy that is increasingly becoming the architecture for transactions, and it’s deeply troubling. And then the last thing I’ll say is, so I’ve already said… like in terms of quantity of jobs, we are already seeing evidence of layoffs, right? We’re already seeing the evidence of layoffs.

It’s just that people aren’t necessarily able to pinpoint and ascribe it to AI. That’s point number one. Two, we need to go beyond the question of quantity of jobs and also look at the impact of this technology on quality of jobs. And third, we need to really deeply think about, again, to Julie’s point, the architectures that can help mitigate some of the potential adverse effects of this technology, both on the quantity and the quality of jobs. And we don’t have the luxury to sit and wait and say, hey, let’s get the empirical evidence and then we’ll figure out what to do. That will be way too late, right? So what do we need? We need countries to think about competition policy.

We need to look. We need to look very closely at tax policy. We need to look very closely at how labor laws need to change. We need to look at social protection systems. We need to look at skill systems, everything that Julie just mentioned, right? But we have to start from an urgency about this is having a huge impact already. It is likely to be, you know, even bigger, and we don’t have the luxury of time to just sit back and wait and say, hey, we need more empirical evidence before we figure out how to mitigate the negative or potentially negative circumstances. So that is what I think is, you know, really, really urgent, that everyone get on that bandwagon and say we need to create these systems and ask for them and do it in our work and do it in our advocacy.

Anurag Behar

Yeah, thank you. I’ll just follow up. I’ll just follow up with it. So, and Julie, please. Pardon. for saying this. I’m saying this tongue in cheek and all my friends and colleagues here who are not from India please pardon me for what I’m going to say. So, you know, we Indians, why should we care about all this? And the reason I’m saying that is because, you know, well just about 9 or 10 % of our employment is in the formal sector. So even if there is huge disruption in labour markets, maybe 2 % of these people are going to lose their jobs, right? So why should we care about all this stuff? Do you have any comments?

Sabina Dewan

I do. You can be sure I do. You can be sure I have a comment about that. So if you look at the numbers, we are more than 90 % in India in formal employment. So Anurag’s exactly right. He knows his numbers. So, you know, essentially what you’re saying is 1 out of every 10 people stands to be potentially affected, right? That’s one way. of looking at it. The other way of looking at it is we have such few good jobs, right? We have such few jobs in the formal labor market. Only one in 10 people get to have a formal sector job. And now you’re taking that away as well, right? That stands to be disrupted.

So again, we’re moving to a world of work that is much more precarious, much more insecure, much more uncertain, where workers don’t, they’re not even called workers anymore. We call them self -employed contractors. They have no health insurance. They have, you know, this is the precaritization of the labor market. So not only do you have, you know, pandemic, climate change, energy transition, trade shocks, and AI destruction, but you have a world of work that is much more precarious, disrupting everything, but you also now are moving to a place where work is becoming more and more informal. Formal jobs are jobs are being, you know, gotten rid of in the name of, please apologize, in the name of efficiency gains, right?

And so, yeah, so that’s why in India we should be really scared because we have such few formal jobs. And then imagine if you have these jobs in the IT sector in Bangalore disappearing, all the workers that used to go to bars and restaurants and get loans to buy houses and cars, that starts to disappear and it has cascading effects across the economy. So, you know, so the impact of this is definitely in the global south. It is definitely beyond the few formal sector jobs. And it’s deeply disturbing. And we need to actually work to understand from technologists very clearly, you know, how these efficiency gains are going to happen and how they’re going to, how.

What can different governments. and so on, and for architecture, public architecture, manage some of these changes. So we do need to care. Definitely need to care. We need to care urgently.

Anurag Behar

All right. So I’m going to come to Julie on this and come back to you, Sandhya, because I put a pin on something that you said, right? So, Julie, I mean, let’s assume that the alarm that Sabina is raising is at least half true, right? It’s more than half. You know, I have a deep conflict of interest, and I’ll tell you once I’m sort of done with this. So, Julie, how can, you know, what are the lessons that you’re seeing across countries, you know? You’re seeing the vast landscape, right, and IDRC has a view across the continents. So what lessons can be learned? From across the continent. such that AI is able to create opportunities, right, part of what Sandhya talked about, and doesn’t really deepen inequality or it minimizes it.

What are you seeing across the countries? Something, some good stuff.

Julie Delahanty

What is that? What is that regulation? And I think one of the – we have this AI – the Global Index on Responsible AI that some of you may have heard about. It’s been talked about a lot during the conference, or at least some of the sessions that I’ve been to. And really what that is, it’s the largest global rights -based data set on responsible AI. And what is distinctive about it is that it includes a dedicated focus on labor protection and the right to work. And by providing that country level, that sort of comparable data, it looks at 138 countries. So by providing that comparable data, it’s helping governments to understand what they might need to do better, what some of the issues are, how they can improve.

So really using that information to support governments in understanding what is the regulation, what is the solution that they need, not just – You know, it has to be based on some evidence. And I think the third big thing, which won’t be a surprise to anybody here that I’m saying this, is that we really need to have good evidence, and evidence really matters when it comes to these issues. So tools like the Global Index on Responsible AI really allows policymakers to move beyond kind of the abstract must -fix regulation to assess how governance of AI actually affects people’s rights, affects their jobs, affects their working conditions, and supports more proactive policymaking on labor regulations, again, skills, social protections, et cetera.

And I think equally important is that we’re still learning. There is no standardized, here is the regulation that you need codified. Through the kind of work that we’re doing, I think we’re learning what’s the balance between… supporting innovation… and still supporting regulation and safety. And I think working together across many countries to share that kind of information is what’s going to support us in finding the right tools.

Anurag Behar

Thanks, Julie. I’m going to come to you, Sandhya. But I just want to disclose something to all of you. That’s my conflict of interest. You know, Sabina is a labor market researcher, and naturally I would think she’s saying what she’s saying. Julie represents IDRC, and therefore she’s saying what she’s saying. Sandhya is the tech person here, so she’s saying what she’s saying. My problem is I’m responsible for this organization, Azeem Premji Foundation. And my problem is the following. My problem is that the foundation owns about 70 % of Wipro. Okay. So whatever is good for a tech company. is good for us, right? On the other hand, my job is not to take care of the technology and this world.

My job is to take care of the most vulnerable people in the country, right? The very poorest, the most marginalized, those who have no recourse to social protection. That’s my job. So I am a deeply conflicted person, right? Very deeply conflicted person. And I wanted to disclose that because I’m going to come to that towards the end. And it has a specific bearing on the question that I’m going to ask Sandhya, which is, you said something fascinating. And I want to put a pin on that. And I’m pulling your leg, you know, which is that rarely do you hear such words from a tech person. She talked about human care and wisdom, right? Didn’t she?

Okay. So, you know, really, my takeaway from what you were saying is that the tech tech stuff, you know, the coding and that kind of stuff, that can get automated. but something that is human understanding people understanding desires how do you work with people that’s what is hard to do and that’s something that you’re already seeing right so would you want to sort of comment on that

Sandhya Ramachandran Arun

yeah so the stereotype of techies aren’t human is a little unfair I think so don’t anchor it in your heads but then yeah so where do I start at the end of the day what does technology consulting and technology services try to do they try to help our client businesses become more successful and our client businesses in turn become more successful when they are innovative when they are creative when they are growing when they are growing when they are making their business and they are doing their business and they are doing their business and they are doing their business profitably and they are doing their business and they are doing their business and they are doing their business and they are doing their business and they are doing their business Or if they have already reached a state of maturity, they are trying to bring in a whole lot of efficiencies as well, right?

So it’s the S curve where you have an idea, you nail it, and then you kind of scale it, and then you kind of start sailing. And when you’re sailing, that’s when you become a big battleship and you have to focus on discipline and efficiency and ensure that you’re making profits just the same even while you’re running this big ship. But then the cycle doesn’t end there. It kind of keeps going. You keep coming up with new ideas, you keep scaling it, and you keep sailing it. And so profitability starts off with an investment, it grows, and then you have to become super efficient to remain profit. And I’m saying this to my boss because every dollar that we earn funds to the tune of about 66 cents whatever efforts the KMG Foundation uses for welfare, right?

And I think it’s a beautiful thing. It’s a beautiful thing. It’s a beautiful thing. model and I don’t think an AI could have thought of it. So therefore I do believe very strongly that creativity, wisdom, vision, foresight, human centricity is core to any technology disruptor that comes about. So if you imagine the days when there were horse carriages all the horses would have been kind of crowding the roads and people would have been going from place to place and at the end of the day you would have had a whole lot of methane which would have kind of ended the year a long time back because of global warming. But yes, vehicles did come and you did have carbon fuel and the evolution continues.

So I don’t think technology is going to stop. So human ingenuity is going to keep bringing technology disruptors. These technology disruptors are going to be more and more exponential in terms of what they can do. And it is up to humans to figure out how to create policy, how to create a governance mechanism, and how to ensure that we derive benefits, mitigate the risks, and at the same time ensure that humanity is at the center of all of this. Right? Now, this is easier said than done, but we’ve done it with nuclear energy. Despite the disasters, the fact that you and I are still alive today and thriving and living a better life than we ever lived in the last 100 years is an example that, yes, you could have accidents that are preventable, but accidents are created by humans.

And it’s up to the leadership to ensure that they put the required guardrails. It could be policy. It could be governance. It could be guidelines, whatever you call it. And you can even hire a leader. And you can even hire a leader. some of the

Anurag Behar

Yeah, it’s good to hear that, you know. I’m just going to come to one round and then perhaps have the last word, if I may. Yeah, okay. So, Sabina, what’s your take? What should we do? What should we do, really?

Sabina Dewan

So I’ve already kind of said what we should do, but first, Sandhya, everything you said really resonated with me, right? And I fully agree that, you know, that the humans have to take responsibility. I can think of a few very worrying scenarios where there are leaders in the world that have access to, you know, nuclear weapons that perhaps… shouldn’t have access to nuclear weapons, right? So how much confidence do we have in people, and particularly when you look at the overall trend of growing precarity? Again, take India alone. Fifty -eight percent of our employment is now self -employment. It is not, you know, and these are people, workers, that have no coverage of health insurance or any kind of safety net.

Add to that the fact that, like, there’s all these different forces coming that we don’t know, you know, if AI disrupts jobs or pandemics happen. We all saw what happened with migrant workers walking back to their villages, hundreds and thousands of migrant workers, right? There is a lot more precarity in the labor market than there ever has been in the past in modern history. And the problem is that regulation, and the regulation of the labor market, and the regulation of the labor market, and the regulation of the labor market, across the globe are getting weaker and weaker in this respect. And then we don’t have precedent, as Julie said. Like, we’re still trying to figure out exactly what we should do, right?

But I will say, I mean, I’ve said many, and I will say that, you know, in the meantime, AI is different because this is also the first time research is now showing, the first time that the current generation of young people have shown cognitive decline, right? So, I mean, rates of depression, rates of anxiety, cognitive decline. How does cognitive decline affect your ability to operate at work and then be replaced by machines that are more efficient because you’re getting stupider? Like, right? Sorry, but this is a really worrying scenario. So what should we do? I think I’ve said this. Multiple times. Regulation and building of social institutions. institutions, but I’ll take Julie’s challenge and say, okay, let’s go a level deeper.

I think we need to look at competition policy very closely. We need to look at antitrust. We need to look at tax, and within tax, we need to look at, you know, how do we do look at the full gamut from, you know, certain kinds of transaction taxes to what person, like a wealth tax, you know, the whole corporate tax rates, the whole gamut of tax tools that we have at our disposal. We certainly, in an area that I know well, need to look at labor regulations, right? There’s a lot of discussion now about what should happen in the gig economy, but, you know, what about, how do we, how do you distinguish if two people have lost their job, how do you distinguish, you know, between them?

You can’t say, okay, this person lost their job to AI, so we’re going to give them health care and, you know, other kinds of support, but… person we’re not right like you need to have universal systems of support for workers of health care of other forms of Social Security that that enable consumption smoothing as well so the economies keep functioning we need to invest heavily in our skill systems for all the talk I can talk about Indian numbers till I’m blue in the face of all the investment and talk about skills training in India only 4 .1 percent of respondents in our labor force survey acknowledge you know identify as having any kind of formal skills only 4 .1 percent despite you know us saying skill India and talking about investments and skills for the last you know well over a decade and a half skill systems.

There’s also well -documented research about how education, you know, the quality of education is so poor. So how do you take a young person in a remote part of India who can barely read and write, might say that I’ve graduated, I’ve done eighth grade or tenth grade, eighth class, tenth class, you know, even twelfth class, but can barely do foundational reading or math, right? How do you take them and say, I’m going to train you for AI. Yeah, that’s what I’m going to do. Like, it doesn’t work. It doesn’t work. So we need to actually fundamentally think about regulations. We need to very urgently work on our education and skill systems that meet people where they are.

We need to definitely think about universal social protection systems. That enable workers to transition between occupations from one sector to another, from one to another. to another from one occupation to another. And I can go into much more detail because this is something that my organization has worked a great deal on. What kind of systems do we need to enable workers to be better protected and be able

Anurag Behar

Thanks, Sabina. We’ve got, I think, five minutes or so, so I’m going to try and wrap up. Judy, would you want to comment?

Julie Delahanty

Yeah, I just want to make a fairly random point, I think. And that is, in addition to the Artificial Intelligence for Development program that we have, we also have a Future of Work project. And I think one of the interesting things there that we don’t talk about as much, everybody is very worried about job loss. That’s kind of the big, it’s job loss. But actually, one of the bigger issues that’s happening is rethinking how to work and ways of working and the disruption that’s happening within jobs and within the workplace. And so I think that’s a really good point. institutions and organizations, that’s not necessarily about job losses. It’s about a complete shift in the way that we do our work and how workers are going to adapt to that fundamental shift in the way that they work.

So it was just a random thought.

Anurag Behar

I don’t think it’s a random thought at all. I think it’s a salient foundational thought, you know, for this discussion. You want to comment on that one line? Because that’s such an important point.

Sabina Dewan

Yeah, no, I mean, just to say that, you know, the Future Works Collective is a global consortium of researchers that IDRC funds that JustJobs is part of that focuses exactly on that. So I agree 100 % that that is a foundational and very important issue.

Anurag Behar

Sandhya, what about you? How would you want to respond to everything Sabina has said?

Sandhya Ramachandran Arun

Look, I think… Watching and waiting is certainly not an option. I mean, we don’t want to be in a Game of Thrones situation when you’re saying winter is coming for some 22 seasons and then it comes. Nobody’s going to wait for it. So we know what’s coming, and we know what’s coming is also capable of evolving and changing tremendously. So we need to learn to change. And yes, we do need to elect good leaders. We do need to have policy at all levels. We need to have policy embedded in platforms. And of course, we need to have a lot of reimagining work and training of workforce. So yes, I think to some extent, painting doom and gloom is good.

Then we start acting, right? But to some extent, I think it also shouldn’t make you paranoid that you become deer in headlights. So yes, we should act, and we should move forward on all of that that all of us agree on.

Anurag Behar

It seems so. It seems so, absolutely. No, but, you know, I think that’s, in some senses, a very good summary, what you just now said, right? What I wanted to say was that this phrase that’s used, boomer and doomer, boomer and doomer. So in a sense, my head is the boomer and my heart is the doomer, given my role. I want to take you just for a minute, which is my job is more to do with education. So we run three universities. We work with, at any point in time, we are working with more than 100 ,000 teachers, right? And so I’m an education person. I’m not the labor market or the tech person here, right?

And I am deeply concerned by the effect of AI on education, deeply, deeply concerned. In fact, I feel that AI is attacking the very foundation of education. The very foundation of education. What AI is doing is saying, the phrase artificial intelligence, it suggests what it does, which means you essentially outsource your thinking. So teachers are outsourcing their thinking and students are outsourcing their thinking. So essentially, and that’s what Sabina was referring to, but she was referring to in the context of social media, that for the first time in this round of sort of assessments, we are seeing cognitive declines, or on test measures we are seeing declines in student performance. I cannot tell you how serious the issue is.

And it’s impossible to regulate this. It’s impossible to regulate this because it’s everywhere. So the only way we are able to deal with this, in the universities at least, is that all assessment, examination, is now returning to the old world paper and pencil, in class test. No home assignments, no project planning, no test. No project work, nothing. Just come here and sit. and write the examination. It is truly serious. I mean, we don’t know how to tackle this right now. And the reason I talk about that is I want to go back to what the analogy that Sandhya used. And I’m so glad that she did that, which is that it is as serious as the nuclear technology.

It is as serious as the nuclear technology. And in one very deep way, it is far more serious than the nuclear technology because nuclear technology did not reach out and affect every individual human being. The possibility of policies and governance to be able to circumscribe, to put boundaries, to manage, those possibilities were far greater. And the possibilities here were highly disruptive, not highly, perhaps the most disruptive of technologies is in retail form, right? This is retail transformation of humanity. It is so hard. to do this. But I’m really glad. I’m glad that with the three of you here, we have this sort of reasonable conclusion, if I may say so, that we are really facing something as serious as the nuclear technology.

And you can’t run away from it. It’s happening. You can’t run away from it. Job losses will happen. We’ve got to figure a way out of it. And I would want to close on this human note, that eventually, perhaps, those jobs that require wisdom, empathy, care, human understanding, they are going to be the hardest to replace if at all. And they will stay. And that’s what one can see in the tech world. So with that, I want to thank all three of you. Thank you so much. I want to thank all of you for coming here. Thank you very much. Thank you.

Sabina Dewan

Speech speed

146 words per minute

Speech length

Inequality and broader socio‑economic effects

Topics

The digital economy | Social and economic development | Artificial intelligence

Anurag Behar

Speech speed

152 words per minute

Speech length

Skills development, reskilling, and the role of human wisdom

Topics

Capacity development | Artificial intelligence

Sandhya Ramachandran Arun

Speech speed

158 words per minute

Speech length

Regulation, governance, and policy response

Topics

Artificial intelligence | The enabling environment for digital development

Julie Delahanty

Speech speed

150 words per minute

Speech length

Skills development, reskilling, and the role of human wisdom

Topics

Artificial intelligence | Social and economic development

Agreements

Agreement points

AI is causing significant transformation in work and employment that requires immediate action

Speakers

– Sabina Dewan
– Sandhya Ramachandran Arun
– Julie Delahanty
– Anurag Behar

Speakers

– Sabina Dewan
– Julie Delahanty
– Sandhya Ramachandran Arun

Arguments

Urgent need for comprehensive policy responses including competition policy, tax policy, labor law reforms, and universal social protection systems

Strong institutions, regulatory frameworks, and research ecosystems are essential for understanding and managing AI’s labor market impacts

We do need to elect good leaders. We do need to have policy at all levels

Summary

All three speakers agree that robust governance structures, policy frameworks, and institutional responses are crucial for managing AI’s societal impacts effectively

Topics

Artificial intelligence | The enabling environment for digital development | Human rights and the ethical dimensions of the information society

Artificial intelligence | The digital economy | Capacity development

Unexpected consensus

The seriousness and urgency of AI’s impact comparable to nuclear technology

Speakers

– Sandhya Ramachandran Arun
– Anurag Behar

Arguments

We’ve done it with nuclear energy. Despite the disasters, the fact that you and I are still alive today and thriving and living a better life than we ever lived in the last 100 years is an example

AI represents a retail transformation of humanity similar to nuclear technology but more pervasive, affecting every individual

Explanation

It’s unexpected that the technology industry representative (Sandhya) would agree with such a serious comparison of AI to nuclear technology, acknowledging both the transformative potential and the significant risks involved

Topics

Artificial intelligence | Human rights and the ethical dimensions of the information society | The enabling environment for digital development

The need for immediate action despite uncertainty about outcomes

Speakers

– Sabina Dewan
– Sandhya Ramachandran Arun
– Julie Delahanty

Arguments

Urgent need for comprehensive policy responses including competition policy, tax policy, labor law reforms, and universal social protection systems

Watching and waiting is certainly not an option

Strong institutions, regulatory frameworks, and research ecosystems are essential for understanding and managing AI’s labor market impacts

Explanation

Despite representing different perspectives (labor advocate, tech industry, development research), all three speakers converge on the need for immediate policy action rather than waiting for more evidence, which is unexpected given their different institutional positions

Topics

Artificial intelligence | The enabling environment for digital development | Human rights and the ethical dimensions of the information society

Overall assessment

Summary

The speakers demonstrate remarkable consensus on the transformative nature of AI and the urgent need for comprehensive responses, despite representing different sectors (labor advocacy, technology industry, development research, and education). Key areas of agreement include: the unprecedented scale of AI’s impact, the inadequacy of waiting for more evidence before acting, the essential role of human qualities in future work, and the critical need for strong governance frameworks.

Consensus level

High level of consensus with significant implications for policy development. The convergence of perspectives across different stakeholders suggests a mature understanding of AI’s challenges and creates a strong foundation for coordinated action. The agreement spans technical, social, economic, and governance dimensions, indicating that comprehensive, multi-stakeholder approaches to AI governance have broad support among experts from diverse backgrounds.

Differences

Artificial intelligence | The enabling environment for digital development | Human rights and the ethical dimensions of the information society

Unexpected differences

Optimism versus pessimism about human agency in AI governance

Speakers

– Sandhya Ramachandran Arun
– Sabina Dewan

Arguments

Technology consulting success depends on creativity, wisdom, vision, and human-centricity that AI cannot replicate

Tech companies are already laying off thousands of workers, with AI being a major disruption factor on top of other economic forces

Explanation

Despite both being experts in their fields, there’s an unexpected philosophical divide – Sandhya expresses confidence in human ability to manage AI transitions and maintain control, while Sabina expresses deep concern about current trends and human capacity to govern effectively, particularly given existing institutional weaknesses

Topics

Artificial intelligence | Human rights and the ethical dimensions of the information society | The enabling environment for digital development

The significance of India’s informal economy in AI impact assessment

Speakers

– Anurag Behar
– Sabina Dewan

Arguments

Jobs requiring human understanding, empathy, care, and wisdom will be hardest to replace and most likely to survive AI disruption

In India, with only 10% formal employment, disrupting these few good jobs would push more workers into precarious informal work

Explanation

Anurag initially suggests India’s large informal economy might make AI disruption less concerning, but Sabina forcefully argues the opposite – that having so few formal jobs makes AI disruption even more devastating. This represents an unexpected disagreement about whether India’s economic structure provides protection or vulnerability

Topics

Artificial intelligence | The digital economy | Closing all digital divides

Overall assessment

Summary

The main disagreements center on the severity and timeline of AI’s labor market impacts, with fundamental differences between optimistic adaptation versus urgent intervention approaches

Disagreement level

Moderate to high disagreement with significant implications – while all speakers acknowledge AI will transform work, their vastly different assessments of urgency, severity, and required responses could lead to very different policy recommendations and preparedness strategies. The disagreements reflect deeper philosophical differences about human agency, institutional capacity, and the nature of technological disruption.

Partial agreements

All speakers agree that some form of governance and institutional response is needed to manage AI’s impacts, but they disagree on the urgency and specific approaches – Sabina calls for immediate comprehensive action, Julie emphasizes building strong institutions and research capabilities, while Sandhya focuses on adaptation and human-centric approaches

Speakers

– Sabina Dewan
– Julie Delahanty
– Sandhya Ramachandran Arun

Arguments

Urgent need for comprehensive policy responses including competition policy, tax policy, labor law reforms, and universal social protection systems

Strong institutions, regulatory frameworks, and research ecosystems are essential for understanding and managing AI’s labor market impacts

Technology consulting success depends on creativity, wisdom, vision, and human-centricity that AI cannot replicate

Topics

Artificial intelligence | The enabling environment for digital development | Governance and Regulatory Responses

Both agree that AI is fundamentally transforming work beyond simple job displacement, but Sabina focuses on the negative impacts like surveillance and inequality, while Julie emphasizes the need for adaptation to new ways of working

Speakers

– Sabina Dewan
– Julie Delahanty

Arguments

AI enables surveillance and influences decisions about work allocation and worker entitlements, exacerbating inequality

AI is changing how work is done within existing jobs, requiring fundamental shifts in workplace adaptation beyond just job losses

Topics

Artificial intelligence | The digital economy | Human rights and the ethical dimensions of the information society

Artificial intelligence | The digital economy | Capacity development

Takeaways

Key takeaways

AI is already causing significant job displacement, with companies privately acknowledging 30-40% productivity gains leading to workforce cuts, despite public denials

The impact extends beyond job quantity to job quality, with AI enabling increased surveillance, algorithmic management, and worker precaritization

AI represents a ‘retail transformation of humanity’ that is more pervasive than previous technologies like nuclear power, affecting every individual

Jobs requiring human wisdom, empathy, care, and understanding will be the hardest to replace and most likely to survive AI disruption

In developing countries like India, disrupting the small formal employment sector (only 10%) would push more workers into precarious informal work

AI is fundamentally attacking education by encouraging outsourcing of thinking, leading to observable cognitive decline in students

The technology requires urgent, comprehensive policy responses rather than a ‘wait and see’ approach, as the impacts are already manifesting

Resolutions and action items

Implement comprehensive policy reforms including competition policy, antitrust measures, and tax policy reforms (wealth tax, transaction taxes)

Reform labor regulations to address gig economy issues and provide universal social protection systems

Invest heavily in education and skill systems that meet people where they are, moving beyond the current inadequate training (only 4.1% of India’s workforce has formal skills)

Develop human-centric AI through co-creation with workers, communities, and employers

Establish strong regulatory institutions and research ecosystems to track real-world AI impacts on labor markets

Create universal social protection systems to enable worker transitions between occupations and sectors

Return to traditional assessment methods in education (paper-pencil, in-class tests) to combat AI-enabled academic dishonesty

Unresolved issues

How to effectively regulate AI’s impact on education when the technology is ubiquitous and impossible to control

How to train workers with poor foundational literacy and numeracy skills for an AI-driven economy

How to distinguish between different causes of job loss for policy purposes (AI vs. other economic factors)

What specific regulatory frameworks will be most effective across different countries and contexts

How to balance innovation support with worker protection and safety regulations

How to address the fundamental conflict between AI’s efficiency gains and employment preservation

How to manage the transition for workers in developing countries with limited formal employment opportunities

Suggested compromises

Transform junior developers into ‘AI managers’ who oversee AI coding while focusing on design, architecture, and security

Shift hiring criteria from pure technical skills to learnability, communication, and adaptability

Use AI to augment rather than replace human workers in fields like healthcare and finance, maintaining human oversight for decision-making

Accept that some job displacement is inevitable while building robust social safety nets and retraining programs

This comment is insightful because it shifts focus from the binary question of job loss/creation to the more nuanced reality of work transformation. It recognizes that even jobs that survive AI will be fundamentally different, requiring new forms of adaptation and support systems.

Impact

This observation broadened the discussion beyond simple displacement metrics to consider the qualitative transformation of work itself. It helped other participants recognize that the challenge isn’t just about protecting jobs, but about preparing workers for entirely new ways of working, adding complexity to the policy solutions being discussed.

Overall assessment

Explanation

She mentioned that job losses in formal sectors like IT would affect workers in bars, restaurants, and other services, but this interconnected impact needs more systematic study and measurement.

Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.